Iron, nutrient, and phytoplankton distributions in Oregon ...avangeen/pdf/Chase_JGR_02.pdf · Low levels of iron can limit the size of phytoplankton blooms and nutrient consumption;

Iron, nutrient, and phytoplankton distributions in Oregon coastal

waters

Zanna Chase,1,2,3 Alexander van Geen,1 P. Michael Kosro,4 John Marra,1

and Patricia A. Wheeler4

Received 18 May 2001; revised 5 December 2001; accepted 14 March 2002; published 25 October 2002.

[1] The relationship between iron and nitrate concentrations was examined off the coast ofOregon during the upwelling season. Surface Fe and N (nitrate + nitrite) concentrationsmeasured underway by flow injection analysis ranged from <0.3 to 20 nmol L�1 and <0.1 to30 mmol L�1, respectively. Total dissolvable Fe concentrations, measured in unfiltered,acidified samples in surface waters and in vertical profiles, ranged from <0.3 to 300 nmolL�1. Surface water Fe and N concentrations were highly variable and uncoupled. Ourobservations indicate two dominant sources of Fe to Oregon coastal waters: Slope or shelfsediments and the Columbia River. Sedimentary iron, probably largely in the particulateform, appears to be added to surface waters through wind-induced vertical mixing duringstrong winds, through thickening of the bottom mixed layer during relaxation ordownwelling favorable wind conditions, and through outcropping of shelf bottom watersduring upwelling events. The existence of multiple iron sources and the generally high ironconcentrations may explain why the distribution of phytoplankton, measured both remotely(by Sea-viewing Wide Field-of-view Sensor) and underway (by in vivo fluorescence),appeared to be driven primarily by physical dynamics and was not strongly linked to thedistribution of iron. Nevertheless, at some offshore stations where underway Feconcentrations were <0.3 nmol L�1, underway measurements of the physiological state ofphytoplankton by fast repetition rate fluorometry were consistent with mild iron stress,and cross-shelf nutrient distributions were consistent with iron regulation of the magnitudeof phytoplankton blooms. INDEX TERMS: 4875 Oceanography: Biological and Chemical: Trace

elements; 4805 Oceanography: Biological and Chemical: Biogeochemical cycles (1615); 4853 Oceanography:

Biological and Chemical: Photosynthesis; 4845 Oceanography: Biological and Chemical: Nutrients and nutrient

cycling; KEYWORDS: iron, Oregon, upwelling, nutrients, coastal, chlorophyll

Citation: Chase, Z., A. van Geen, P. M. Kosro, J. Marra, and P. A. Wheeler, Iron, nutrient, and phytoplankton distributions in Oregon

coastal waters, J. Geophys. Res., 107(C10), 3174, doi:10.1029/2001JC000987, 2002.

1. Introduction

[2] Because of their proximity to shelf sediments andterrestrial sources, coastal waters are typically enriched iniron relative to the open ocean [Johnson et al., 1997]. Yetrecent work in the California coastal upwelling system[Hutchins and Bruland, 1998; Hutchins et al., 1998] hasdemonstrated that the addition of iron to incubation bottlesfrom some coastal regions stimulates phytoplankton growthand nitrate consumption much as it does in open oceanhigh-nutrient, low-chlorophyll (HNLC) regions [e.g., Coaleet al., 1996]. Thus, iron is potentially an important variable

controlling phytoplankton biomass and community compo-sition [Bruland et al., 2001; Johnson et al., 2001] inproductive coastal upwelling systems.[3] Most of the work to date on iron in coastal upwelling

systems has focused on waters off central California. Dis-solved iron concentrations in upwelling source waters(depth of �150 m, �175 km offshore) of the CaliforniaCurrent are relatively low (<1 nmol L�1) [Martin andGordon, 1988], which suggests that, in contrast to its effecton the macronutrients, upwelling of deep water does notsignificantly enrich surface waters in iron. Additionalpotential iron sources to the coastal ocean include shelfsediments, wind-blown dust, and riverine inputs. Given therelatively high precipitation north of central California, andthe prevailing alongshore wind direction, aeolian inputs ofiron are expected to be small. The concentration of dis-solved iron in river water can be several orders of magni-tude greater than open ocean concentrations [Boyle et al.,1977]. However up to 95% of riverine filterable iron can belost through estuarine processes such as flocculation andprecipitation [Boyle et al., 1977; Mayer, 1982a], so evenlarge rivers may contribute very little dissolved iron to the

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. C10, 3174, doi:10.1029/2001JC000987, 2002

1Lamont-Doherty Earth Observatory of Columbia University, Palisades,New York, USA.

2Also at Department of Earth and Environmental Sciences, ColumbiaUniversity, Palisades, New York, USA.

3Now at the Monterey Bay Aquarium Research Institute, MossLanding, California, USA.

4College of Oceanic and Atmospheric Sciences, Oregon StateUniversity, Corvallis, Oregon, USA.

Copyright 2002 by the American Geophysical Union.0148-0227/02/2001JC000987$09.00

38 - 1

coastal ocean. Indeed, resuspended shelf sediments werefound to be the dominant source of Fe to surface waters offcentral California, even during periods of increased riverrunoff [Johnson et al., 1999].[4] The west coast of North America has considerable

alongshore variability in upwelling characteristics, climateand shelf geometry, all of which could affect Fe inputs. Forinstance, the continental shelf off Oregon, defined by the200 m isobath, varies from a minimum width of 11 km, justoff Cape Blanco (42�N) to a maximum of 78 km, at HecetaBank (44.2�N), and is generally wider than the shelf offCalifornia. Furthermore, in the summer, surface waters offOregon are strongly influenced by freshwater input from theColumbia River, which has its maximum discharge in June,when the plume is directed almost exclusively southwardover the Oregon shelf.[5] This paper describes high-resolution surface, and

bottle-based vertical distributions of key chemical and bio-logical properties measured off southern Oregon and north-ern California in July 1999. The objective of this study wasto understand the physical processes governing iron inputsduring the upwelling season, in order to better understandthe development of HNLC conditions in the coastal ocean.The specific hypotheses pertaining to this region include: 1)Low levels of iron can limit the size of phytoplanktonblooms and nutrient consumption; 2) Shelf sediments arean important source of Fe to surface waters; 3) TheColumbia River is an important source of Fe to surfacewaters and 4) Iron and macronutrients are input to surfacewaters by different mechanisms and at different times.

2. Methods

2.1. Sample Collection

[6] Sampling took place off the Oregon and Californiacoasts between July 3 and July 9, 1999, aboard the R/VWecoma. Time, position, bottom depth, wind speed,downwelling solar radiation (285–2800 nm), flow throughsea surface temperature (SST) and flow through salinitywere recorded as 1 min averages by the ship’s data logger(XMIDAS). Shipboard ADCP data were collected andprocessed as described by Kosro [2002]. When the shipwas underway at 10 knots, a continuous stream of near-surface (�1–3 m depth), trace metal clean seawater wasobtained for iron and nitrate analyses using a brass ’fish’ (abathythermograh) towed over the side of the ship [Boyle etal., 1982; van Geen et al., 2000; Vink et al., 2000]. The fishwas suspended from a boom about 5 m off the port side, andits torpedo shape helped maintain the tubing intake pointingforward and into the water. Through its �1 min transit fromthe ocean to a class 100 laminar flow bench, the pumpedseawater was only in contact with Teflon1-lined tubing,with the exception of a �50 cm length of silicone tubing(9.5 mm ID; Masterflex1) in a large peristaltic pump. Alltubing was leached with 1.2 N HCl for at least 24 h and thenrinsed thoroughly with ultrapure MQ water (18.2 M�,Millipore) prior to use. The pumped flow was normallydirected into a 50 mL overflowing graduated cylinderwithin the flow bench, from which a flow for underwayiron and nitrate measurements was drawn by a peristalticpump. By switching acid-cleaned Teflon1 valves this flowwas occasionally diverted to fill archive bottles for future

analysis. The pumping system was turned off during CTDintensive transects. These lines were in most cases imme-diately resampled after the CTD casts as a continuoustransect for underway measurements.[7] Profile samples for Fe analysis were collected in acid-

leached high-density polypropylene bottles from Niskinbottles (General Oceanics) modified for trace metal sam-pling by acid leaching, replacing inner springs with C flextubing, and viton o rings with fluorosilicone O rings. Astandard metal rosette frame and hydrowire were used. Alldiscrete samples, including those from the underway pump-ing system, were acidified at sea by the addition of 1 mLL�1 of concentrated Seastar HCl and analyzed within 7months by the flow injection method described below.

2.2. Flow Injection Analysis

[8] Nitrate and iron were detected colorimetrically by flowinjection analysis using a three-channel fiber-optic spectro-photometer (Ocean Optics). The sensitivity of both methodswas increased with long path length spectrophotometric cells(LPC) using the Teflon AF-2400 (Biogeneral) liquid corewaveguide approach described byWaterbury et al. [1997]. A10-port injection valve (VICI) was used to control sampleand reagent flow, which permitted both the Fe and NO3

sample loops to be filled simultaneously. A 10-port selectionvalve (Cheminert) was used to direct either standards or thesample stream into the system, with at least 4 standards runtypically every hour. Surface seawater samples were injectedevery 80 seconds. Standards, run in triplicate, were mixedsolutions of iron and cleaned (Chelex-100) nitrate, acidifiedwith 1 mL L�1 HCl as described above for discrete samples.Nitrate was undetectable in the iron stock solution and ironwas undetectable in the cleaned nitrate stock solution.2.2.1. Iron[9] Iron was measured using a modification of the method

of Measures et al. [1995], with a 10 cm LPC. We did notpreconcentrate the samples, in order to reduce sampling time,and thereby maximize the spatial resolution of sampling, andto avoid cross contamination in this region of strong gra-dients in surface Fe concentrations. To further simplify thesystem and minimize contamination, the sample stream wasneither filtered nor acidified before injection. The lack offiltration caused clogging of the valves on several occasionsin nearshore waters, which lead to system downtime. Unac-idified MQ water (UA-MQ) was run with each set of stand-ards. Usually, the absorbance of the UA-MQ was less thanthat of acidified MQ water, presumably because acidifiedMQ was leaching some adsorbed Fe from the system. Sincethe seawater stream was unacidified, we subtracted the UA-MQ peak from every sample peak as a blank value and tookthe sensitivity (absorbance per mole Fe) from the slope of thestandard curve using acidified standards. Most of the timethe UA-MQ blank was not detectable above the baseline. Thedetection limit was �0.3 nmol L�1, based on variability ofthe blank. The precision, based on triplicate analyses ofstandards, was 1–6%. The response of the detector for Fewas linear up to a concentration of 100 nmol L�1. Drift overan hour separating standards, both in terms of blank andsensitivity, was generally <20%, and was assumed to belinear as a function of time.[10] The same flow injection system was used for analyz-

ing discrete samples for iron. Five of the ports on the 10-

38 - 2 CHASE ET AL.: IRON IN OREGON WATERS

port selection valve were dedicated to samples, and flushedthoroughly between runs. Samples containing Fe above thelinear range of the method were diluted with acidified MQand rerun.[11] The current chemical Fe literature is confusing

because of the use of different terminology resulting fromthe wide variety of methods used to analyze Fe, and thesimilarly wide variety of preanalysis sample treatmentsemployed. In this paper we discuss both discrete samples,which were acidified with 2 mL L�1 12N HCl and stored atleast 2 months prior to analysis, and underway samples,unacidified and analyzed inline at a reaction pH of 6.4. Atthe low pH (�2) of the unfiltered samples, organicallybound Fe, colloidal Fe, and some, if not all, particulate Fewould certainly be released into solution during storage. Wewill therefore refer to the discrete analyses as ’dissolvableFe’ (dFe). This probably represents the maximum poten-tially bioavailable iron in the system over biologicallyrelevant timescales. Our terminology is similar to thatemployed by Bruland and Rue, [2001] when referring tothe analysis of unfiltered, acidified, stored, samples. Thisshould not be confused with the ‘‘DI Fe’’ used by Martinand Gordon [1988] to denote dissolved (0.4 mm) Fe. Notethat the term ‘‘dissolvable Fe’’ has also been used todescribe shipboard Fe measurements by FIA of unfilteredsamples following a brief period of acidification to pH �3[Obata et al., 1997; Johnson et al., 1999]. This ‘‘dFe’’ iscloser to the ‘‘underway Fe’’ described in this paper,although we omitted the brief acidification step.[12] If we can be reasonably confident that acidified,

stored samples return something close to the ’total’ (dis-solved + particulate) Fe in a water sample, the fraction andnature of Fe detected by other methods, including theunderway method used here, is far less certain and is oftensimply operationally defined. The Fe that we measuredunderway in an unacidified, unfiltered stream, at a reactionpH of 6.4 for �40 seconds most likely includes only themost reactive Fe, and therefore represents a minimumestimate of bioavailable Fe. Preliminary experiments usinga large excess of EDTA as a chelator suggest organicallybound Fe is not detected under these conditions. However,tests against EDTA-complexed Fe may be misleading, sincethese complexes are not typical of naturally occurringiron binding ligands, which tend to have lower concentra-tions and lower conditional stability constants than EDTA[Morel and Hering, 1993]. Despite this acknowledgeduncertainty in the nature of the iron detected by the under-way measurements, we believe they are still informative,particularly in showing regional differences within thisstudy area. Unfortunately, differences in methodology makeit difficult to compare our results with those of otherworkers using different methods. Even the work of Vinket al. [2000], whose methodology most closely resemblesours, is not directly comparable, because inline acidificationwas used in the Vink et al. study (as by Johnson et al.[1999]), and in addition their samples were preconcentratedbefore injection.2.2.2. Macronutrients[13] Nitrate + nitrite (N) was measured underway, as

described above using the standard azo dye colorimetricmethod [e.g., Anderson, 1979]. The detection limit was 0.1mmol L�1 and the precision 1–6%. The response was linear

to 100 mmol L�1. Nitrate (N), phosphate (P), and silicic acid(Si) in surface waters were also measured in the laboratoryin discrete samples collected from the underway pumpingsystem using a commercial flow injection system (LachatQuikChem 8000) with standard colorimetric methods, as byvan Geen et al. [2000]. Profile samples for macronutrientanalyses were collected from the CTD rosette and wereneither filtered, nor acidified, and were stored frozen.Macronutrients from the CTD profiles samples were ana-lyzed by standard wet chemical methods according toprotocols of Gordon et al. [1995] using an Alpkem RFA-300.

2.3. Fluorometry

[14] In vivo fluorescence was measured underway fromthe ship’s flow though intake (�3 m depth) with a TurnerDesigns (model 10-AU) fluorometer. The in vivo fluores-cence signal was calibrated with extracted chlorophyllsampled from the instrument’s outlet and extracted chlor-ophyll from surface and 10 m Rosette samples at CTDstations, interpolated to 3 m.[15] The fast repetition rate fluorometer (Chelsea Instru-

ments) was run in underway mode, from the same samplestream as the underway fluorometer. We used a flashsequence of 100 saturation flashes of 1.1 ms duration,separated by �2.8 ms followed by 20 relaxation flashes of1.1 ms duration, separated by 51.6 ms. Each sample was anaverage of 10 sequences. A constant gain of 1 was usedthroughout the cruise. Optical surfaces were cleaned daily.Data were reduced using the Chelsea Instruments softwareto obtain the initial, or minimal (Fo) and maximal (Fm)fluorescence yield. From these parameters we derive theratio Fv/Fm = (Fo-Fm)/Fm, which is the maximum changein the quantum yield of fluorescence, a quantitative measureof the efficiency of photosystem II.

3. Results

3.1. Physical Setting

[16] An image from the Advanced Very High ResolutionRadiometer (AVHRR) on July 3 shows cool water nearshorethroughout the cloud-free study area, with the coldest waterssouth of Cape Blanco (43�N; Figure 1). An 8 day compositeimage from the Sea-viewing Wide Field-of-view Sensor(SeaWiFS; data obtained through the Goddard DAAC athttp://daac.gsfc.nasa.gov/data/data set/SEAWIFS/)(Figure 1b) shows a visual association between cool SSTin the AVHRR image and high phytoplankton biomass,despite the difference in temporal coverage between thetwo images (Figure 1). South of 42�N both SST andchlorophyll occurred in distinct plumes. Many of the featuresof the SSTand chlorophyll distributions, including a seawardextension of cool, chlorophyll-rich water south of CapeBlanco (Figure 1), can be related to the acoustic Dopplervelocity at 17 m. The surface velocity field defined thetrajectory of a coastal jet, which can be followed from45�N, where it was present nearshore, to �44.2�N, where itbegan to separate from the coast, then returned shoreward at43�N, and finally moved offshore south of Cape Blanco(41�N), similar to the flow described by Barth et al. [2000]for a previous year. The flow south of 42�N is particularlyenergetic and complex, and is consistent with the intrusion of

CHASE ET AL.: IRON IN OREGON WATERS 38 - 3

relatively warm, low-chlorophyll waters near the shelf break,evident in the AVHRR and SeaWiFS images (Figure 1).[17] Wind vectors from three stations within the study

area were broadly coherent (by inspection), and demonstratethe dominance of upwelling favorable winds (equatorward,negative) during this season (Figure 2). However, windrelaxation events did occur, for example at the beginning ofthe cruise (07/03–07/04), on 07/07 north of 43�N, and on07/09 at 41�N (Figures 2a and 2b).

3.2. Regional Pattern of Near-Surface Properties

[18] Cape Blanco defined a boundary between two distinctregimes in SST, surface salinity and surface chlorophyll a(Figure 1 and Figures 3a, 3b, and 3c). Waters to the north ofCape Blanco, seaward of the 50 m isobath, were warm (up to17�C), fresh (salinity as low as 23 psu) and containedrelatively little chlorophyll (<2 mg L�1). Upwelling centers

of cool SST, high salinity and elevated chlorophyll wereobserved near the HH line (44�N) and north of the NH line(44.7�N), but thesewere restricted towithin the 200m isobath(Figure 3). South ofCapeBlanco, very cold (9�C), salty (up to34 psu) and chlorophyll-rich (�20 mg L�1) waters wereobserved along the coast. Cool (<13�C), salty (up to 33.5psu), chlorophyll-rich (5–10 mg L�1) waters extended sea-ward of the 1000m isobath in this region. The very cold SSTsand high salinity measured near the CR line on 07/05 are aclear indication of strong upwelling. Winds recorded to thenorth and south of the CR line (42�N) at this time were indeedupwelling favorable (Figures 2b and 2c), but they were notparticularly strong. This apparent discrepancy between windand upwelling strength, as inferred from tracer distributions,can be attributed to the enhanced shoaling of the pycnocline,for a given wind stress, as a current flows cyclonically arounda prominent feature, such as Cape Blanco [Arthur, 1965].

Figure 1. ADCP velocity vectors (longest arrow represents 42 cm/s) at 17 m between 3 and 8 July 1999overlaid on satellite-derived (a) sea surface temperature (�C) from 5 July and (b) an 8 day (7/4–7/11)composite of satellite-derived chlorophyll a concentration (mg L�1). Chlorophyll is displayed on alogarithmic scale. There are no chlorophyll data in the black regions near the coast. A probable path ofthe coastal current has been subjectively drawn through the velocity vectors in Figure 1a. The locations ofthree transects where high-resolution data are examined more closely are also indicated in Figure 1a.Initials along the coast indicate the standard sampling lines of the Oregon GLOBEC program. See colorversion of this figure at back of this issue.


[19] High concentrations of N tended to be associatedwith nearshore, cold, high-salinity waters (Figures 3 and 4),consistent with a primary supply from upwelling. In generalthere was good agreement between underway and discretemeasurements of N (r2 = 0.83; n = 65; Figures 4a, 4b, and5a). The generally higher values in discrete samples couldbe due to acid solubilization of particulate organic nitrogenincluded in the unfiltered discrete samples; some of thelargest outliers were indeed from regions of high chloro-phyll (Figure 5a). Some of the scatter in this relationshipmay also be due to smearing of signals in the underwaysystem and the difficulty of precisely time matching under-way and discrete samples, especially in regions of stronghorizontal gradients in N.[20] The distribution of Si, in contrast to N, was not well

delimited by Cape Blanco. Si-rich waters were present bothsouth of Cape Blanco, in the cold, salty water along the coastnear the CR line, and also north of Cape Blanco, in warm,fresh waters as far as the 1000 m isobath (Figure 3d). Thepresence of the Columbia River plume was indicated by thevery low salinity (23–24 psu), high Si (up to 41 mmol L�1)and warmer SSTs [Park et al., 1972] measured offshore atthe northern end of the cruise track. A riverine source of Si,in addition to the upwelling source, probably explains whythe distribution of Si was poorly delimited by Cape Blanco.[21] Of the underway observations in this study, 55%,

were of Fe < 1 nmol L�1, and of samples containing morethan 10 mmol L�1 N (n = 105), 20% had Fe < 0.5 nmol L�1.

Iron concentrations measured in stored, acidified samplescollected from the underway pumping system were up to20-fold higher than underway measurements by FIA (Fig-ures 4c, 4d, and 5b). The difference between the twomeasurements was greatest in shallow, nearshore waters(Figures 4c and 4d), which are more likely to containelevated levels of particulate Fe [Johnson et al., 1997],and possibly colloidal Fe, which would have been partly orcompletely solubilized by acidification. Underway Fe con-centrations were not well correlated with either salinity(Figure 6b), N (Figure 6c) or SST (not shown).[22] The distribution of Fe resembled the distribution of Si

more so than the distribution of N: High Fe was observed incold, salty waters along the coast south of Cape Blanco, andalso in low salinity waters of the Columbia River plume,particularly along 45�N (Figures 4c and 4d). Considering allof the data together, discrete Fe concentrations were bettercorrelated with Si than with N or P (Figure 7). Phosphateconcentrations closely followed nitrate (Figure 7f ), with aslight excess of P over N, as observed in other coastalupwelling systems [Brzezinski et al., 1997]. The poorercorrelation between Si and N and P, compared to the strongcorrelation between N and P (Figures 7b, 7d, and 7f ) againmost likely reflects the fact that Si has both an upwelling anda riverine source.

3.3. Vertical Distributions

[23] Selected profiles of density, N, chlorophyll a and dFeare shown in Figure 8. In each case, an offshore station, theupwelling source region, is paired with a nearshore stationalong the same sampling line. All measurements of Fe invertical profiles were from discrete samples (i.e., totaldissolvable iron, dFe). In general, less than 20 hoursseparated sampling the offshore and the nearshore stations,except along the NH line, where the offshore station wassampled 4 days before the nearshore station.[24] The distinction between waters to the north and south

of Cape Blanco is visible in the offshore density profiles,with waters north of Cape Blanco being far more stratifiedthan those to the south (CR and EU lines) (Figure 8). Thedistribution of nitrate was generally related to density,presumably through wind-driven shoaling of the isopycnalsurfaces. The densest water (sq = 26) and highest nitrate(>20 mmol L�1) was found in the deepest water at allstations, and the stations with highest surface density (NHand CR, nearshore) had the greatest surface nitrate (Figure8). Low surface nitrate at the nearshore EU station, despitehigh density, suggests some consumption of N hadoccurred.[25] In contrast, the distribution of iron was not obviously

related to density (Figure 8). As expected, offshore dFeconcentrations were generally lower than nearshore concen-trations at all depths. Although concentrations near thedetection limit were measured just below the surface onthe NH and FM lines, dFe concentrations greater than 10nmol L�1 were observed at depth at the offshore stations onall but the NH line, with dFe > 30 nmol L�1 on the HH, CR,and HH lines (Figure 8). These ‘‘offshore’’ dFe concen-trations are an order of magnitude greater than thosemeasured a similar distance from shore off Monterey Bay[Martin and Gordon, 1988; Johnson et al., 1999]. It isdifficult to imagine that methodological differences, for

-

Figure 2. Wind vectors (m s�1) measured from (a) nearthe NH line at the Newport buoy (44.6�N, 124.5�W; NOAAbuoy 46050), (b) near the FM line at the Cape Aragolighthouse (43.4�N, 124.4�W; Station CARO3), and (c)near the EU line at the St George buoy (41.0�N, 124.4�W;NOAA buoy 46027). The duration of the R/V Wecomacruise is indicated by a thin line in all panels, and theapproximate time of sampling along GLOBEC lines in thevicinity of each wind station is indicated by initials alongthis line. The time of the detailed transects in the vicinity ofeach wind station (TR1, TR2, and TR3; see Figure 1) isindicated by a bold line and numeral in each panel. Notethat TR1, along the CR line, is located between the FM andthe EU lines and TR2 is between the NH and FM lines.Wind data are available from the National Data Buoy Center(http://www.ndbc.noaa.gov/Maps/Northwest.shtml).


example, in the strength and duration of acidification, couldaccount for such a large difference between the two regions.[26] Phytoplankton biomass, which was included in the

unfiltered water samples used for Fe analysis, accounted fora small fraction of the total Fe measured in most profilesamples. For example, assuming at the high end, as much as50 mgC:mgchla [Parsons et al., 1977] and as much as 100mmolFe:molC [Sunda and Huntsman, 1995] for coastalphytoplankton, 6 mg L�1 of chlorophyll a yields at mostonly 5 nmol L�1 Fe. Thus, while phytoplankton biomasscould account for most of the Fe in surface waters from theoffshore stations (Figure 8), it can account for at most 3% ofthe high dFe concentrations found nearshore, or at depth atsome of the offshore stations (e.g., HH, CR, and EU).[27] At two nearshore stations along the NH line, dFe and

beam attenuation (particle concentration) had similar profileshapes, both with a middepth minimum (Figure 9), suggest-

ing an association between the Fe and the particulate matter.Themost obvious explanation is that the Fe is associated withlithogenic particulate matter resuspended from sediments[Martin and Gordon, 1988; Johnson et al., 1999]. Althoughchlorophyll (fluorescence) does not contribute significantlyto dFe loads (see discussion above), it does contribute tobeam attenuation [Small and Curl, 1968]. The presence ofchlorophyll at the surface therefore complicates the relation-ships between beam attenuation, particulate matter, and Fe.We have used the relationship between beam attenuation andfluorescence, derived from all CTD casts (not shown), tocorrect beam attenuation for the attenuation associated withfluorescence. When corrected for chlorophyll, the beamattenuation signal is reduced at the surface, but not signifi-cantly at depth (Figure 9), and the middepth minimum inbeam attenuation is preserved. This suggests that the pro-nounced middepth minimum in dFe is indeed linked to the

Figure 3. Surface water properties measured off Oregon during July 3–8, 1999. (a) Temperature (�C),(b) salinity, and (c) chlorophyll a (mg L�1, estimated from in vivo fluorescence) were measuredcontinuously while the ship was underway. (d) Silicate (silicic acid) was measured in discrete, acid-preserved samples in the laboratory. Where transects were covered twice during the cruise, the later dataare plotted over the earlier data. Note the logarithmic color scale used for chlorophyll a and silicateconcentrations. Contour lines represent the 50, 200, and 1000 m isobaths. See color version of this figureat back of this issue.


depth distribution of lithogenic particles. Note that althoughthe two stations had approximately the same ‘‘nonchloro-phyll’’ particle content near the surface, near-surface dFe wasa factor of 2 lower at 9 km from shore compared to 5 km fromshore (Figure 9). This may reflect either a seaward decreasein the concentration of nonparticulate Fe, or a seawarddecrease in the Fe content of particulate matter.

3.4. High-Resolution Cross-Shelf Transects

[28] High-resolution, underway mapping is ideally suitedto the coastal upwelling environment, where small-scalespatial variability is common [Chavez et al., 1991]. Here wepresent three cross-shelf transects of underway data that coverthe sampled area and a range of upwelling conditions (TR1,

TR2, and TR3; see Figures 1 and 3 for location). Discretebottle data from these transects are reported in Table 1.[29] Winds were upwelling favorable at the time of sam-

pling TR1 (Figure 2), and the low SST, high N (up to 15 mmolL�1) and high salinity nearshore indicate active upwelling(Figure 10). However, iron concentrations measured under-way were low (<0.5 nmol L�1). Maximum chlorophyll wasfound 15 km offshore, coincident with a region of lower N,but with no corresponding change in SST or salinity (Figure10). From the difference between the observed N concen-tration (6 mmol L�1), and that predicted based on the moreconservative tracers, SST and salinity (12 mmol L�1), weestimate a removal of about 6 mmol L�1 N by phytoplanktongrowth.

Figure 4. Surface water concentrations of (a) and (b) nitrate + nitrite (N) (mmol L�1) and (c) and (d)iron (nmol L�1) off Oregon, July 3–8, 1999, measured in discrete, acidified (pH 2) samples collectedfrom the underway pumping system or (unacidified) surface rosette samples (Figures 4a and 4c) and atsea by flow injection analysis, without filtration or acidification (Figures 4b and 4d). Note the logarithmiccolor scale. Gray circles in Figure 4c indicate the location of CTD stations where vertical profiles of Nand Fe were obtained. See color version of this figure at back of this issue.


[30] Transects TR2 and TR3 were sampled duringupwelling favorable winds following about 2 days ofdownwelling (Figure 2a). Nearshore SSTs along TR2 wererelatively warm (14�C) compared to SSTs further south(Figure 10), and surface water nitrate was undetectablewithin about 20 km from shore (Figure 10). Thus, thesurface waters at TR2 had not yet been modified after�12 hours of upwelling favorable local winds (Figures 6aand 6b), which is consistent with previous work showing ittakes about 1 day for the upwelling circulation off Oregonto respond to the wind [Barber and Smith, 1981; Brink etal., 1994]. However, in contrast to TR1, Fe concentrationsalong TR2 were high within 15 km from shore. Along TR2there was a distinct cool (11–13�C) patch, enriched in bothN and Fe (5–8 mmol L�1 N and �15 nmol L�1 Fe) locatedbetween 20 and 30 km from shore (Figure 10). Inspection of

the ADCP velocity field at 17 m and an AVHRR imagefrom July 3 (Figure 1a) suggests this cold feature may havebeen advected southwestward from the region of cool SSTand high N nearshore on the NH line (Figure 3c).[31] Waters along TR3 were influenced by the Colum-

bia River plume, as seen by their low salinity (<31 psu;Figure 10), and high Si (Table 1). Along TR3, in contrast toTR2, decreasing SSTs and increasing salinity, N and Fetoward the coast indicate surface waters had begun to bemodified by upwelling. The contrast between the cross-shelfdistribution of SST, salinity and N along TR1 and TR3suggests upwelling was more intense (faster and fromgreater depth) at TR1 than at TR3, although winds weremore strongly equatorward at TR3 than at TR1 at the time ofsampling (Figure 6). Underway Fe was about 10 timesgreater all along TR3 compared to TR1, while dFe was

Figure 5. A comparison for (a) N and (b) Fe between measurements made underway by flow injectionanalysis on an unacidified sample stream and measurements from discrete bottle samples collected fromthe underway stream and stored acidified at pH 2 for 7 months prior to analysis. The1:1 line is drawn forreference. Note the semilog scale in Figure 5b. Samples with chlorophyll a > 10 mg L�1 are in solidcircles.

Figure 6. Scatterplot of surface salinity versus underway (a) nitrate + nitrite (N) and (b) Fe; (c)scatterplot of underway N versus underway Fe.


greater nearshore at TR1, but greater offshore at TR3 (Figure10 and Table 1). A small-scale anomaly was observed alongTR3 in the temperature and salinity data between 45 and 50km from shore, with an enrichment of Fe and N observed atthe seaward edge of this feature (Figure 10).

3.5. Fast Repetition Rate Fluorometry

[32] The photosynthetic competency parameter, Fv/Fm,displayed strong diel variability, with high values at night andlow values during the day. This behavior has been observedbefore [e.g., Behrenfeld and Kolber, 1999] and is most likelycaused by nonphotochemical quenching at high light levels.At irradiances below 500 W m�2 d�1, Fv/Fm decreasedroughly linearly with increasing light, but there was too muchscatter in this relationship for it to be used to light correct theFv/Fm data. We therefore restrict our analysis to FRRfmeasurements made between 21:00 and 05:00.[33] The ratio Fv/Fm reaches a maximum value of 0.65

under optimal conditions, and deviations from this ratio havebeen interpreted as symptomatic of physiological stress, mostnotably nutrient limitation [Greene et al., 1994]. In thepresent study, nighttime Fv/Fm varied between 0.4 and 0.6,with a mean of 0.48, suggesting a range from nutrient-repleteto mildly nutrient-stressed phytoplankton. A similar rangewas observed on a transect between the Sargasso Sea andDelaware Bay [Geider et al., 1993a], while values of 0.25 aretypical of the Fe-limited equatorial Pacific [Behrenfeld et al.,1996; Behrenfeld and Kolber, 1999]. Although there was atrend to lower values away from the coast, at least south of45�N (Figure 10 and Figure 11d), overall, no correlation wasfound between Fv/Fm and ambient nutrient concentrations,or between Fv/Fm and temperature or salinity (not shown).The lowest values of nighttime Fv/Fm in this study (�0.4)were measured at stations CR7, CR8 and FM8 (Table 2).

Both dFe and Si concentrations were relatively low at thesestations and the low Fv/Fm might reflect either mild Fe stressor mild Si stress (or some other stressor). Experiments withphytoplankton cultures suggest the ratios of maximal andminimum fluorescence to chlorophyll a, Fo/chl and Fm/chl,in dark adapted cells should increase under Fe and NO3

limitation [Geider et al., 1993b], decrease under Si limi-tation [Lippemeier et al., 1999], and be unaffected by Plimitation [Geider et al., 1993b]. We have evaluated theseparameters at all sites where nighttime Fv/Fm measure-ments were made at CTD stations, enabling a matchbetween average continuous FRRf parameters (while onstation) and chlorophyll a measured in the surface Rosettesample (Table 2). Although the data are limited, the ratiosFo/chl and Fm/chl at CR7 and FM8 were the highest of anyof the stations, which suggests the low Fv/Fm observed atthese stations is unlikely to be due to Si limitation [Lippe-meier et al., 1999]. At FM8, the low N and (relatively) lowdFe, low Fv/Fm and high Fo/chl are consistent with both Nand Fe nutrient stress [Geider et al., 1993b].

4. Discussion

[34] Iron and nitrate concentrations off Oregon displayedcomplex spatial patterns in all dimensions. These distribu-tions are clearly influenced by multiple factors, includingtopography, shelf geometry, and the magnitude and varia-bility (spatial and temporal) of wind forcing. The followingdiscussion focuses on mechanisms of iron inputs to theOregon coastal region, and their biological implications.

4.1. Iron Sources

[35] The upwelling origin of surface water nitrate is wellknown from previous studies of coastal upwelling systems

Figure 7. Scatterplots of nutrient-nutrient and iron-nutrient relationships from discrete (acidified) bottlesamples collected from the underway surface water pumping system. The correlation coefficient (r) isshown on each plot (n = 122 samples).


[e.g., Chavez et al., 1991; Hayward and Venrick, 1998] andcan be seen here, for example, in the similarity betweendensity and nitrate profiles (e.g., Figure 8). Along allsampling lines, N levels at the nearshore station were lessthan or equal to the highest concentrations found in the top200 m at the offshore station (Figure 8), which is consistentwith an offshore, deep-water origin for nearshore N. Incontrast to N, surface water underway Fe and nearshore dFeat all depths along most lines were higher than in offshoreprofiles, confirming the existence of iron sources distinctfrom the upwelling that delivers N to surface waters [Martin

and Gordon, 1988]. Here we evaluate three sources of Fe toOregon surface waters: shelf sediments, Fe-enrichedupwelling source waters, and the Columbia River, with anemphasis on how iron inputs relate to wind forcing andtherefore nitrate supply (Table 3). We have no evidence tosuggest a significant input of Fe from atmospheric sources,but note that they may be important locally, particularly nearindustrial centers [Hardy et al., 1985].4.1.1. Shelf Sediments[36] Shelf sediments have been identified as the main

source of Fe to the upwelling region off central California

Figure 8. Vertical profiles of density anomaly (sq; kg m�3), nitrate + nitrite (N) (mmol L�1), chlorophyll

a (mg L�1), and dissolvable iron (dFe) (nmol L�1) at the offshore (open symbols and thin lines) andnearshore (solid symbols, thick lines) ends of the lines sampled in July 1999. The offshore dFe profilesare replotted in the last panel to enlarge the scale. Station lines are presented top to bottom from north tosouth, with station locations shown in Figure 3c. On the NH and HH lines, chlorophyll a was estimatedfrom in vivo fluorescence using the relationship between fluorescence and extracted chlorophyll aderived from all other casts. In general, less than 20 hours separated sampling the offshore and thenearshore stations, except along the NH line, where the offshore station was sampled 4 days before thenearshore station.


[Johnson et al., 1999]. The importance of the shelf insupplying iron to California coastal waters is highlightedby the finding that north of Monterey Bay, where thecontinental shelf extends 20–50 km from shore, surfacewaters are iron-replete, whereas off the Big Sur coast, wherethe shelf extends only several km, surface waters are iron-depleted [Bruland et al., 2001]. The high Fe concentrationsfound in this study over the broad shelf region of HecetaBank are an indication of the importance of sedimentary Fesources off Oregon. However, the mechanism by whichsediment-derived Fe is introduced to surface waters is notwell understood. Based on our observations, we proposethree such mechanisms: (1) wind mixing, (2) outcropping ofthe bottom boundary layer, and (3) thickening of the bottommixed layer.[37] Input of sedimentary Fe to surface waters through

bottom-reaching, wind-induced vertical mixing is likely to beimportant only in nonstratified water columns, very near-shore, and at high wind speeds. This mode of Fe input will becoupled to N input only when strong winds are upwellingfavorable. Such conditions may have occurred nearshore atthe CR line in July 1999, where a profile of dFe showed alinear increase with depth, suggesting a sedimentary source.Although winds at the time of sampling were not extremely

strong (�7 m s�1 from the ship and the CARO3 lighthouse),the wind stress may have been sufficient to resuspend bottomsediments and their associated particulate and dissolved Febecause the water column was not stratified. The very lowsurface water Fe measured underway at this site, in contrastto the high dFe and N values, suggests either the mixingprocess supplied only particulate Fe from sediments, orconsumption and scavenging of dissolved Fe was very rapid.[38] Another way for sediment and surface water to

interact is if the thickness of the bottom mixed layer(BML, a vertically homogeneous layer above the bottom)increases until it intersects the surface. Sediment-rich (andtherefore, presumably, iron rich), nepheloid layers are oftenassociated with BMLs. Observations from the northernCalifornia and Oregon shelves have shown that suppressednear-bottom vertical mixing, and hence thin BMLs, prevailduring upwelling conditions, while enhanced vertical mix-ing and thick BMLs prevail during downwelling conditions[Pak and Zaneveld, 1977; Lentz and Trowbridge, 1991]. ABML develops rapidly following the onset of downwellingconditions [Pak and Zaneveld, 1977; Lentz and Trowbridge,1991], which means even short-term wind reversals maylead to resuspension of Fe. The frequency of wind reversalevents may therefore be a critical parameter in determining

Figure 9. Vertical profiles of dissolvable iron (dFe) (nmol L�1), fluorescence (Volts), beam attenuation(m�1), and (d) beam attenuation corrected for chlorophyll, from stations at 9 km (thin line, open symbols)and 5 km (thick line, solid symbols) from shore along the NH line.

Table 1. Discrete Nutrient Analyses From Along the High-Resolution Underway Transects Presented in

Figure 10

Transect Latitude, �N Longitude, �W Distance, km P, mmol L�1 Si, mmol L�1 N, mmol L�1 dFe, nmol L�1

TR141.90 �124.65 49.7 0.4 1.2 0.1 <0.341.90 �124.58 41.9 0.5 <0.2 2.6 <0.341.90 �124.53 35.4 0.6 <0.2 3.3 <0.341.90 �124.43 25.2 1.1 10.9 11.3 0.741.90 �124.38 19.1 1.1 18.5 11.1 58.441.90 �124.32 12.2 1.6 32.0 18.1 117.0

TR244.01 �124.19 6.1 0.3 2.5 <0.1 194.944.05 �124.24 11.6 0.3 0.2 0.3 27.544.09 �124.29 17.5 0.3 4.5 1.4 13.244.13 �124.33 22.9 0.7 10.3 6.3 9.344.19 �124.41 32.6 0.1 1.9 <0.1 44.0

TR344.87 �124.28 26.4 0.13 6.37 1.49 28.044.73 �124.52 50.9 0.10 5.88 <0.1 31.944.65 �124.65 72.2 0.10 8.35 <0.1 36.1


Fe inputs in coastal upwelling systems. This mechanism ofiron input also helps explain the apparent decoupling betweenN and Fe distributions (e.g., Figure 6c), since it provides Feinput during downwelling, precisely when N inputs areminimal. Temporal variability in the height of the BML couldalso provide an explanation for themiddepthminimum in dFeand particle concentration profiles observed at nearshorestations on the NH line at the onset of upwelling conditions(e.g., Figure 9). It is conceivable that during downwellingconditions, which preceded sampling at theseNHstations, theBML thickened and the entire water column became turbidand iron-rich. With the onset of upwelling, the middepthminimum could then have been generated as an intermediatelayer of clear, iron-poor water penetrated onto the shelf [Pakand Zaneveld, 1977; Lentz and Trowbridge, 1991].[39] In the absence of vertical mixing, diffusion of Fe

from sediments will reach at most several cm above the

sediment-water interface [Kremling and Petersen, 1978],and will not be easily entrained in the onshore flow.However, recent models [Allen et al., 1995] and observa-tions [van Geen et al., 2000; Takesue and van Geen, 2002],suggest a substantial portion of upwelling flow occurs in thebottom boundary layer (BBL) and outcrops in very shallowwaters near the coast. Therefore, even over steep topogra-phy (narrow shelf), some portion of an N-rich upwellingplume always interacts with the shelf where it may becomeenriched in Fe. Outcropping of the BBL may thereforeexplain the high dFe concentrations measured very near-shore on the FM line, in N-rich, recently upwelled waters.This mechanism could provide an alternative explanationfor middepth minimum observed in the dFe profiles at shelfstations along the NH line at the onset of upwelling (e.g.,Figure 9). Under this scenario, high-Fe waters outcroppedwithin several km from shore during intense upwelling of

Figure 10. Transects TR1 (solid circles), TR2 (open circles), and TR3 (solid squares) showing surfacewater properties measured underway in July 1999: Chlorophyll a (mg L�1) was derived from in vivofluorescence. For locations, see dark bars in Figures 1 and 3. Analyses from discrete samples along thesetransects are presented in Table 1.


the BBL, and were then transported offshore in the surfaceEkman flow, over relatively less Fe-rich waters, to producea middepth minimum in dFe.4.1.2. Enrichment of Offshore Waters[40] The work of Martin and Gordon [1988] off central

California suggests that in the absence of shelf interaction,upwelling of deep offshore water cannot supply sufficientFe to surface waters to support full consumption of theconcurrently upwelled N, even if all forms of Fe areassumed to be bioavailable. Off Oregon, in contrast, wefind dFe >25 nmol L�1 in 4 out of 5 profiles from stationsseaward of the 1000 m isobath. This is about twice theamount of Fe required to support full consumption of the�25 mmol L�1 N present at the depth of upwelling in thesesource waters. The origin of these elevated levels of subsur-face dFe off Oregon is not clear, and warrants further study.At the FM, EU and CR lines, the increases in dFe at depthare associated with slight increases in beam attenuation

(data not shown), which suggests the dFe may be largelyparticulate, and associated with resuspended sediment trans-ported off the shelf [e.g., Pak et al., 1980]. Anotherpossibility is the transport of dissolved Fe from pore watersin fluidized flow along isopycnals as the result of sedimentslumping [Martin and Gordon, 1988].[41] Regardless of its origin, iron enrichment at middepth

(�100 m) in offshore waters clearly represents an importantiron source to nearshore surface waters in this region, viaupwelling. This subsurface iron may also be mixed intosurface waters on the outer shelf by a number of physicalprocesses [e.g., Klein and Coste, 1984; Dewey and Moum,1990; Paduan and Niiler, 1990; Chavez et al., 1991],operating on perhaps very small spatial scales. The smallanomaly observed in Fe and N along TR3 may have beenthe result of such a process, where N and Fe were brought tothe surface by enhanced vertical mixing associated with theSST and salinity front, perhaps in response to wind forcing

Figure 11. The cross-shelf distribution of (a) N (measured underway), (b) Fe (measured underway), (c)dFe (measured in discrete acidified samples), and (d) underway Fv/Fmmeasured at night (between 2100 and0500) during the July 1999 survey. Data from four latitudinal bands are distinguished by different symbols.


across the front [Klein and Coste, 1984; Chavez et al., 1991;Franks and Walstad, 1997].4.1.3. Columbia River[42] The presence of the Columbia River may account for

the high iron levels off Oregon, compared to centralCalifornia. The average discharge from the Columbia Riverat the Dalles, Oregon, in July 1999 was 7112 m3 s�1 (http://waterdata.usgs.gov), which is over 10 times the discharge ofthe largest river in California (the Sacramento) at the sametime (630 m3 s�1; California Department of Water Resour-ces). The high dFe observed as far as the 1000 m isobath inlow-salinity, high-silicate water along 45�N is a strongindication that iron from the Columbia River is transportedhundreds of kilometers downstream. We have not mademeasurements of Fe in the Columbia River itself, but aUSGS program reported dissolved Fe at �200 nmol L�1 inJuly at 3 sites about 140 km upstream from the mouth[Fuhrer et al., 1996], while earlier measurements of dis-solved Fe in the Columbia River, compiled by Riedel et al.[1984], range from 9 to 2800 nmol L�1. More data areneeded, but it appears iron in the Columbia River estuarymay exhibit the near-conservative behavior found in someestuaries with rapid flushing [Mayer, 1982b], rather thanbeing lost through coagulation [Sholkovitz et al., 1978].[43] Iron input from the Columbia River plume is impor-

tant because it represents a source of iron to waters seawardof the shelf break. Furthermore, the Fe at 45�N associatedwith the Columbia River plume is probably in a smaller sizefraction than Fe associated with resuspended sediment, sincelarger particles would have settled out of the plume during its�20 day [Huyer, 1983] transit from the river mouth. This isconsistent with the high levels of Fe measured underway inthe Columbia River plume, since our underway FIA systemwould probably not detect iron associated with large (>col-loidal) particles. The presence of the Columbia River may

also contribute to the shelf sediment source of Fe byexporting Fe-rich particles accumulated in the estuary[Mayer, 1982b; Johnson et al., 2001], or by contributingto rapid sediment accumulation and hence reducing con-ditions conducive to Fe flux from the sediments [Trefry andPresley, 1982]. The contrast between deep dFe concentra-tions offshore of California and Oregon may be due in part tothe presence of the Columbia River, which produces anextensive band of fine sediment on the shelf [e.g., Gross etal., 1967]. Composite SeaWiFS images for July 1998 andJuly 1999 (not shown), both show high phytoplanktonbiomass near the mouth of the Columbia River whichsuggests the river, which provides iron, silicic acid and astable surface layer, may play an important role in maintain-ing biological productivity over the Oregon shelf.

4.2. Biological Implications

[44] This study was undertaken in part to evaluatewhether iron might be a limiting nutrient off Oregon, as itcan be in waters off central California [Hutchins and Bru-land, 1998; Hutchins et al., 1998]. Our data suggest it is not,except possibly on the outer shelf. Compared to open oceanenvironments and the HNLC waters observed off centralCalifornia [Hutchins and Bruland, 1998; Hutchins et al.,1998], iron concentrations, both underway and dissolvable,were high off Oregon. At these nanomolar concentrationsFe is unlikely to be rate limiting, even for coastal phyto-plankton [Sunda and Huntsman, 1995]. Another manifes-tation of iron limitation is Liebig limitation, in which thesize of a phytoplankton bloom is limited by the concen-tration of the least available nutrient, relative to demand.The amount of Fe needed to support full consumption ofupwelled nitrate varies strongly as a function of the ambientiron concentration [Bruland et al., 2001], and this makes itdifficult to predict the Fe:N (or Fe:P) input ratio at which

Table 3. Proposed Sources and Mechanisms of Iron Inputs to Surface Waters and Their Relationship to Nitrate Input

Iron Source Mechanism Winds Nitrate Input

Shelf sediments wind-induced mixing strong; equatorward or poleward Nothickening of bottom mixed layer/bottom nepheloid layer weak or poleward Nooutcropping of the bottom boundary layer strong; equatorward Yes

Offshore deepwater isopycnal shoaling; diapycnal vertical mixing strong; equatorward YesColumbia River plume transport any small

Table 2. Nutrient and FRRf Data From CTD Stations Occupied at Nighta

Station Latitude, �N Longitude, �W Chl a Si P N dFe Fo/chl Fm/chl Fv/Fm

CR 7 41.90 �125.00 2.08 <0.2 0.5 3.0 1.5 276.6 462.5 0.40CR 8 41.90 �125.20 <0.2 0.7 2.7 3.1 0.40CR 9 41.90 �125.33 4.95 2.2 0.8 8.0 3.9 134.6 280.1 0.52FM 5 43.22 �124.67 0.77 3.0 <0.1 <0.1 8.4 184.8 363.3 0.49FM 6 43.22 �124.75 0.55FM 7 43.22 �124.83 0.40 3.0 <0.1 <0.1 3.2 208.0 364.6 0.43FM 8 43.22 �125.00 0.32 0.2 <0.1 <0.1 1.9 240.8 391.2 0.39HH 3 44.00 �124.60 3.5 0.1 <0.1 19.6 0.51HH 4 44.00 �124.80 0.71 2.9 <0.1 <0.1 10.7 132.3 248.6 0.47HH 5 44.00 �125.00 0.90 3.0 <0.1 <0.1 4.1 106.8 233.2 0.54HH 6 44.00 �125.10 0.55aSurface chlorophyll a (mg L�1; extracted, not estimated from in vivo fluorescence), macronutrients (mmol L�1), dissolvable iron (dFe and nmol L�1),

and photosynthetic parameters (measured by the underway FRRf). FRRf data acquired every 4 min were averaged over each station occupation. The dFevalue at all stations except CR7 and CR9 is from discrete surface samples collected at station locations during underway transits.


Liebig limitation by iron applies. A further complication isthat a one-time measurement of Fe and N at a given locationsays nothing about the initial ratio at which these nutrientswere added to the system.[45] The best way to assess the nutrient limitation status

of the phytoplankton community may be with nutrientaddition/incubation experiments. However, fast repetitionrate fluorometry (FRRf ) can be used to assess (and map)the physiological status of the phytoplankton community.Overall, the relatively high (0.4–0.6) nighttime values ofFv/Fm found during this study are not consistent withsevere (rate) nutrient limitation of any kind. It also appearsunlikely that iron input was Liebig-limiting the develop-ment of phytoplankton blooms off Oregon in July 1999.We note that the largest bloom observed in this study,south of Cape Blanco, was not coincident with the regionof greatest iron inputs (over and north of Heceta Bank),but rather with the region of strongest upwelling (lowestSST, strongest shoaling of the pycnocline).[46] Another approach to assess nutrient limitation is to

consider the time course of an upwelled parcel of water[Bruland et al., 2001]. For example, along TR1, on the CRline, underway iron fell to below our detection limit closerto shore than did nitrate, at roughly the same distance fromshore that the chlorophyll concentration fell to low levels.One interpretation of these observations is that as thenearshore phytoplankton bloom aged and was advectedoffshore in the surface Ekman layer, it came to an endabout 25 km from shore due to lack of iron. Indeed, theFRRf data (low Fv/Fm and high Fo/chl) from offshorewaters along the CR line were consistent with mild ironstress. The low concentrations of Si observed in thisregion, relative to N, may even be the result of anincreased Si:N uptake ratio in Fe-stressed diatoms [Hutch-ins and Bruland, 1998; Takeda, 1998]. Thus, despite therelatively high iron concentrations in offshore waters ofthe Oregon upwelling region, our analysis of Fv/Fm andFo/chl at CTD stations, and the cross-shelf distribution ofFe,N and chlorophyll does suggest iron can exert somephysiological control on the phytoplankton community. Inthe hierarchy of Hutchins et al. [1998] these may be type2, or ‘‘slightly Fe-stressed’’, communities.[47] From the difference between Fe concentrations

measured underway and in acidified samples, and fromthe coherence between profiles of particle concentrationand dFe, we conclude that much of the iron in the nearshoreenvironment is associated with suspended particles. Thismeans iron concentrations in the euphotic zone will dependcritically on relative rates of solubilization of Fe fromparticles [e.g., Wells and Mayer, 1991; Johnson et al.,1994; Barbeau et al., 1996; Maranger et al., 1998] andsinking rates of particulate Fe in the offshore flowingsurface layer. The observation that the same intensity offluorescence-corrected beam attenuation was associatedwith a higher concentration of dFe nearshore than offshore(e.g., Figure 9) may indicate ’stripping’ of Fe from partic-ulate phases as they move across the shelf [Johnson et al.,1997]. Barber and Smith [Barber and Smith, 1981] haveproposed that the two-layered flow of upwelling systemsmakes them important sites for recycled production. Thesame may be true for Fe, in the sense that any particulate Fewhich sinks into the BBL can be entrained in the onshore

flow and upwelled again to the surface where it may besubjected to further solubilization.[48] The strong baroclinic jets of the Coastal Transition

Zone (CTZ) off California do not transport a significantamount of coastally upwelled nitrate to the ocean interior[Chavez et al., 1991], and this also appears to be the caseoff Oregon. In July 1999, N enrichment did not extendfurther than �50 km from the coast (Figure 10 and Figure11a), even south of Cape Blanco, where the coastal jetmoves offshore [Barth et al., 2000]. Throughout thesurvey area, high dFe levels were generally restricted closeto shore, but elevated levels of Fe were measured under-way at all distances from shore (Figure 10 and Figures 11band 11c), especially north of Cape Blanco, perhaps reflect-ing the diversity (and abundance) of iron sources. Thepotential for Fe export from the coastal ocean probablydepends on the form of nearshore Fe. Iron that has beensolubilized from lithogenic particles [Johnson et al., 1997]and incorporated into the dissolved or biogenic pool maybe more likely to escape sedimentation over the shelf, andbe exported to the ocean interior.

5. Conclusions

[49] The high-resolution surface water transects and verti-cal profiles of chemical and biological parameters presentedhere provide insight into the magnitude, mechanisms andconsequences of iron and nitrate inputs to surface waters ofthe Oregon upwelling system. Our results support conclu-sions for central California that shelf sediments are animportant source of Fe to coastal upwelling systems [Johnsonet al., 1999]. Sedimentary iron, largely in the particulateform, is added to surface waters through wind-inducedvertical mixing, thickening of the bottom mixed layer, andupwelling through the bottom boundary layer. Off Oregon,iron input associated with the Columbia River is also impor-tant. A full description of Fe inputs to the Oregon upwellingsystem will require a better understanding of the magnitude,reactivity and variability of Fe inputs from the ColumbiaRiver. Middepth waters off the Oregon shelf appear tocontain significantly more total Fe than waters off centralCalifornia, and this offshore enrichment also contributes Feto surface waters. As a result of large riverine inputs, arelatively wide shelf and offshore enrichment of deep waters,iron concentrations off Oregon, both labile and total, aregenerally high, and not likely to be limiting biologicalproduction. However, relatively high-nitrate (>10 mmol L�1

NO3), low-iron (<0.5 nmol L�1) waters do exist off Oregon,as they do off California [Hutchins and Bruland, 1998;Hutchins et al., 1998], and this may influence phytoplanktonphysiology and productivity seaward of the shelf break. Thesource and mechanism of iron input is expected to influencethe form of iron in surface waters, its decoupling frommacronutrients, and its residence time. Future work shouldexamine seasonal changes in iron inputs and possible ironlimitation along the Oregon coast [as was done recently forMonterey Bay Johnson et al., 2001] as well as characterizethe biological availability of iron associated with the variousinput mechanisms described in this paper.

[50] Acknowledgments. We thank the officers and crew of the R/VWecoma, who went to great lengths to assist with the underway operations.


N. Ventriello and C. Ho acquired and reduced the FRRf data. AVHRR datawere provided by P.T. Strub (O.S.U.) under funding by the U.S. GLOBECprogram. J. Fleischbein kindly provided the CTD data. Discussions with J.Barth, K. Bruland, B. Hales, R. Letelier, and P. Strutton were very helpful.This work was funded by the Office of Naval Research (NOOO14-99-1-0279 to AvG and JM) and by a fellowship (awarded to ZC) from the Fondspour la Formation de Chercheurs et l’Aide a la Recherche (Quebec). This isLDEO publication number 6315.

ReferencesAllen, J. S., P. A. Newberger, and J. Federiuk, Upwelling circulation on theOregon continental shelf, 1, Response to idealized forcing, J. Phys. Ocea-nogr., 25, 1843–1866, 1995.

Anderson, L., Simultaneous spectrophotometric determination of nitrite andnitrate by flow injection analysis, Anal. Chim. Acta, 110, 123–128, 1979.

Arthur, R. S., On the calculation of vertical motion in eastern boundarycurrents from determinations of horizontal motion, J. Geophys. Res., 70,2799–2803, 1965.

Barbeau, K., J. W. Moffett, D. A. Caron, P. L. Croot, and D. L. Erdner, Roleof protozoan grazing in relieving iron limitation of phytoplankton, Nat-ure, 380, 61–64, 1996.

Barber, R. T., and R. L. Smith, Coastal upwelling ecosystems, in Analysis ofMarine Ecosystems, edited by A. R. Longhurst, pp. 31–68, Academic,San Diego, Calif., 1981.

Barth, J. A., S. D. Pierce, and R. L. Smith, A separating coastal upwellingjet at Cape Blanco, Oregon and its connection to the California CurrentSystem, Deep Sea Res., Part II, 47, 783–810, 2000.

Behrenfeld, M. J., and Z. S. Kolber, Widespread iron limitation of phyto-plankton in the South Pacific Ocean, Science, 283, 840–843, 1999.

Behrenfeld, M. J., A. J. Bale, Z. S. Kolber, J. Aiken, and P. G. Falkowski,Confirmation of iron limitation of phytoplankton photosynthesis in theequatorial Pacific Ocean, Nature, 383, 508–511, 1996.

Boyle, E. A., J. M. Edmond, and E. R. Sholkovitz, The mechanism of ironremoval in estuaries, Geochim. Cosmochim. Acta, 41, 1313–1324, 1977.

Boyle, E. A., S. S. Huested, and B. Grant, The chemical mass balance of theAmazon plume, II, Copper, nickel and cadmium, Deep Sea Res., Part 1,29, 1355–1364, 1982.

Brink, K. H., J. H. Lacasce, and J. D. Irish, The effect of short-scale windvariations on shelf currents, J. Geophys. Res., 99, 3305–3315, 1994.

Bruland, K. W., and E. L. Rue, Iron: Analytical methods for the determina-tion of concentrations and speciation, in The Biogeochemistry of Iron inSeawater, edited by K. A. Hunt and D. R. Turner, pp. 255–289, chap. 6,John Wiley, New York, 2001.

Bruland, K. W., E. L. Rue, and G. J. Smith, Iron and macronutrients inCalifornia coastal upwelling regimes: Implications for diatom blooms,Limnol. Oceanogr., 46, 1661–1674, 2001.

Brzezinski, M. A., D. R. Philips, F. P. Chavez, G. E. Friederich, and R. C.Dugdale, Silica production in the Monterey, California upwelling system,Limnol. Oceanogr., 42, 1694–1705, 1997.

Chavez, F. P., R. T. Barber, P. M. Kosro, A. Huyer, S. R. Ramp, T. P.Stanton, and B. R. Demendiola, Horizontal transport and the distributionof nutrients in the Coastal Transition Zone off northern California: Effectson primary production, phytoplankton biomass and species composition,J. Geophys. Res., 96, 14,833–14,848, 1991.

Coale, K. H., et al., A massive phytoplankton bloom induced by an eco-system-scale iron fertilization experiment in the equatorial Pacific Ocean,Nature, 383, 495–501, 1996.

Dewey, R. K., and J. N. Moum, Enhancement of fronts by vertical mixing,J. Geophys. Res., 95, 9433–9445, 1990.

Franks, P. J. S., and L. J. Walstad, Phytoplankton patches at fronts: Amodel of formation and response to wind events, J. Mar. Res., 55, 1–29, 1997.

Fuhrer, G., G. Tanner, J. Morace, S. McKenzie, and K. Skach, Wateranalysis of the lower Columbia River basin: Analysis of current andhistorical water-quality data through 1994, USGS Water-Resour. Invest.95-4294, U. S. Geol. Surv., Portland, Oreg., 1996.

Geider, R. J., R. M. Greene, Z. Kolber, H. L. Macintyre, and P. G. Falk-owski, Fluorescence assessment of the maximum quantum efficiency ofphotosynthesis in the western North Atlantic, Deep Sea Res., Part I, 40,1205–1224, 1993a.

Geider, R. J., J. Laroche, R. M. Greene, and M. Olaizola, Response of thephotosynthetic apparatus of Phaeodactylum tricornutum (Bacillariophy-ceae) to nitrate, phosphate, or iron starvation, J. Phycol., 29, 755–766,1993b.

Gordon, L. I., J. C. Jennings, A. R. Ross, and J. M. Krest, A suggestedprotocol for continuous flow automated analysis of seawater nutrients(phosphate, nitrate, nitrite, and silicic acid) in the WOCE hydrographicprogram and the Joint Global Ocean Fluxes Study, Rep. 93-1, OregonState Univ., Corvallis, 1995.

Greene, R. M., Z. S. Kolber, D. G. Swift, N. W. Tindale, and P. G.Falkowski, Physiological limitation of phytoplankton photosynthesisin the eastern equatorial Pacific determined from variability in thequantum yield of fluorescence, Limnol. Oceanogr., 39, 1061–1074,1994.

Gross, M. G., D. A. McManus, and H.-Y. Ling, Continental shelf sediment,northwestern United States, J. Sediment Petrol., 37, 790–795, 1967.

Hardy, J. T., C. W. Apts, E. A. Crecelius, and N. S. Bloom, Sea-surfacemicrolayer metal enrichments in an urban and rural bay, Estuarine Coast-al Shelf Sci., 20, 299–312, 1985.

Hayward, T. L., and E. L. Venrick, Nearsurface pattern in the CaliforniaCurrent: Coupling between physical and biological structure, Deep SeaRes., Part II, 45, 1617–1638, 1998.

Hutchins, D. A., and K. W. Bruland, Iron-limited diatom growth and Si: Nuptake ratios in a coastal upwelling regime, Nature, 393, 561–564, 1998.

Hutchins, D. A., G. R. DiTullio, Y. Zhang, and K. W. Bruland, An ironlimitation mosaic in the California upwelling regime, Limnol. Oceanogr.,43, 1037–1054, 1998.

Huyer, A., Coastal upwelling in the California Current system, Prog. Ocea-nogr., 12, 259–284, 1983.

Johnson, K. S., K. H. Coale, V. A. Elrod, and N. W. Tindale, Iron photo-chemistry in seawater from the equatorial Pacific, Mar. Chem., 46, 319–334, 1994.

Johnson, K. S., R. M. Gordon, and K. H. Coale, What controls dissolvediron concentrations in the world ocean?,Mar. Chem., 57, 137–161, 1997.

Johnson, K. S., F. P. Chavez, and G. E. Friederich, Continental-shelf sedi-ment as a primary source of iron for coastal phytoplankton, Nature, 398,697–700, 1999.

Johnson, K., F. P. Chavez, V. A. Elrod, S. E. Fitzwater, J. T. Pennington, K. R.Buck, and P. M.Walz, The annual cycle of iron and the biological responsein central California coastal waters, Geophys. Res. Lett., 28, 1247–1251,2001.

Klein, P., and B. Coste, Effects of wind-stress variability on nutrient trans-port into the mixed layer, Deep Sea Res., Part A, 31, 21–37, 1984.

Kosro, P. M., A poleward jet and an equatorial undercurrent observed offOregon and northern California, during the 1997–98 El Nino, Prog.Oceanogr, in press, 2002.

Kremling, K., and H. Petersen, Distribution of Mn, Fe, Zn, Cd and Cu inBaltic seawater—A study on the basis of one anchor station, Mar. Chem.,6, 155–170, 1978.

Lentz, S. J., and J. H. Trowbridge, The bottom boundary-layer over thenorthern California shelf, J. Phys. Oceanogr., 21, 1186–1201, 1991.

Lippemeier, S., P. Hartig, and F. Colijn, Direct impact of silicate on thephotosynthetic performance of the diatom Thalassiosira weissflogii as-sessed by on- and off- line PAM fluorescence measurements, J. PlanktonRes., 21, 269–283, 1999.

Maranger, R., D. F. Bird, and N. M. Price, Iron acquisition by photosyn-thetic marine phytoplankton from ingested bacteria, Nature, 396, 248–251, 1998.

Martin, J. H., and R. M. Gordon, Northeast Pacific iron distributions inrelation to phytoplankton productivity, Deep Sea Res., Part A, 35, 177–196, 1988.

Mayer, L. M., Aggregation of colloidal iron during estuarine mixing: Ki-netics, mechanism and seasonality, Geochim. Cosmochim. Acta, 46,2527–2535, 1982a.

Mayer, L. M., Retention of riverine iron in estuaries, Geochim. Cosmochim.Acta, 46, 1003–1009, 1982b.

Measures, C. I., J. Yuan, and J. A. Resing, Determination of iron in sea-water by flow injection analysis using in-line preconcentration and spec-trophotometric detection, Mar. Chem., 50, 3–12, 1995.

Morel, F. M. M., J. G. and Hering, Principles of Aquatic Chemistry, JohnWiley, New York, 1993.

Obata, H., H. Karatani, M. Matsui, and E. Nakayama, Fundamantal studiesfor chemical speciation of iron in seawater with an improved analyticalmethod, Mar. Chem., 56, 97–106, 1997.

Paduan, J. D., and P. P. Niiler, A Lagrangian description of motion innorthern California coastal transition filaments, J. Geophys. Res., 95,18,095–18,109, 1990.

Pak, H., and J. R. V. Zaneveld, Bottom nepheloid layers and bottom mixedlayers observed on the continental-shelf off Oregon, J. Geophys. Res., 82,3921–3931, 1977.

Pak, H., J. R. V. Zaneveld, and J. Kitchen, Intermediate nepheloid layersobserved off Oregon and Washington, J. Geophys. Res., 85, 6697–6708,1980.

Park, K., C. Osterberg, and W. Forster, Chemical budget of the ColumbiaRiver, in The Columbia River Estuary and Adjacent Ocean Waters, editedby A. Pruter and D. Alverson. pp. 123–135, Univ. of Wash. Press,Seattle, 1992.

Parsons, T. R., M. Takahashi, and B. Hargrave, Biological OceanographicProcesses, 332 pp., Pergamon, New York, 1977.


Riedel, G. F., S. L. Wilson, and R. L. Holton, Trace metals in the ColumbiaRiver estuary following the 18 May 1980 eruption of Mount St. Helens,Pac. Sci., 38, 340–348, 1984.

Sholkovitz, E. R., E. A. Boyle, and N. B. Price, The removal of dissolvedhumic acids and iron during estuarine mixing, Earth Planet. Sci. Lett., 40,130–136, 1978.

Small, L., and H. Curl, The relative contribution of particulate chlorophylland river tripton to the extinction of light off the coast of Oregon, Limnol.Oceanogr., 13, 84–91, 1968.

Sunda, W. G., and S. A. Huntsman, Iron uptake and growth limitation inoceanic and coastal phytoplankton, Mar. Chem., 50, 189–206, 1995.

Takeda, S., Influence of iron availability on nutrient consumption ratio ofdiatoms in oceanic waters, Nature, 393, 774–777, 1998.

Takesue, R. K., and A. van Geen, Nearshore circulation during upwellinginferred from the distribution of dissolved cadmium off the Oregon coast,Limnol. Oceanogr., 47, 176–185, 2002.

Trefry, J. H., and B. J. Presley, Manganese fluxes from Mississippi deltasediments, Geochim. Cosmochim. Acta, 46, 1715–1726, 1982.

van Geen, A., R. K. Takesue, J. Goddard, T. Takahashi, J. A. Barth, and R. L.Smith, Carbon and nutrient dynamics during coastal upwelling off CapeBlanco, Oregon, Deep Sea Res., Part II, 47, 975–1002, 2000.

Vink, S., E. A. Boyle, C. I. Measures, and J. Yuan, Automated high resolu-tion determination of the trace elements iron and aluminum in the surfaceocean using a towed fish coupled to flow injection analysis, Deep SeaRes., Part 1, 47, 1141–1156, 2000.

Waterbury, R. D., W. S. Yao, and R. H. Byrne, Long pathlength absorbancespectroscopy: Trace analysis of Fe(II) using a 4.5m liquid core wave-guide, Anal. Chim. Acta, 357, 99–102, 1997.

Wells, M. L., and L. M. Mayer, Variations in the chemical lability of iron inestuarine, coastal and shelf waters and its implications for phytoplankton,Mar. Chem., 32, 195–210, 1991.

��Z. Chase, Monterey Bay Aquarium Research Institute, 7700 Sandholdt

Road, Moss Landing, CA 95039, USA. ([email protected])P. M. Kosro and P. A. Wheeler, College of Oceanic and Atmospheric

Sciences, Oregon State University, Corvallis, OR 97331–5503, USA.J. Marra and A. van Geen, Lamont-Doherty Earth Observatory of

Columbia University, P. O. Box 1000, Palisades, NY 10964, USA.


Figure 1. ADCP velocity vectors (longest arrow represents 42 cm/s) at 17 m between 3 and 8 July 1999overlaid on satellite-derived (a) sea surface temperature (�C) from 5 July and (b) an 8 day (7/4–7/11)composite of satellite-derived chlorophyll a concentration (mg L�1). Chlorophyll is displayed on alogarithmic scale. There are no chlorophyll data in the black regions near the coast. A probable path ofthe coastal current has been subjectively drawn through the velocity vectors in Figure 1a. The locations ofthree transects where high-resolution data are examined more closely are also indicated in Figure 1a.Initials along the coast indicate the standard sampling lines of the Oregon GLOBEC program.

38 - 4

CHASE ET AL.: IRON IN OREGON WATERS

o

o

o

o

o

o

o

o

o

o

Figure 3. Surface water properties measured off Oregon during July 3–8, 1999. (a) Temperature (�C),(b) salinity (psu), and (c) chlorophyll a (mg L�1, estimated from in vivo fluorescence) were measuredcontinuously while the ship was underway. (d) Silicate (silicic acid) was measured in discrete, acid-preserved samples in the laboratory. Where transects were covered twice during the cruise, the later dataare plotted over the earlier data. Note the logarithmic color scale used for chlorophyll a and silicateconcentrations. Contour lines represent the 50, 200, and 1000 m isobaths.

38 - 6


Figure 4. Surface water concentrations of (a) and (b) nitrate + nitrite (N) (mmol L�1) and (c) and (d)iron (nmol L�1) off Oregon, July 3–8, 1999, measured in discrete, acidified (pH 2) samples collectedfrom the underway pumping system or (unacidified) surface rosette samples (Figures 4a and 4c) and atsea by flow injection analysis, without filtration or acidification (Figures 4b and 4d). Note the logarithmiccolor scale. Gray circles in Figure 4c indicate the location of CTD stations where vertical profiles of Nand Fe were obtained.

38 - 7


Documents

Iron, nutrient, and phytoplankton distributions in Oregon ...avangeen/pdf/Chase_JGR_02.pdf · Low levels of iron can limit the size of phytoplankton blooms and nutrient consumption;